Open shunlinlu opened 2 years ago
Hi,
do you mean the pre-processing for the dataset? For different datasets (OpenImages and Visual Genome) we have different scripts. It is somehow complex. I recommend that you follow the COCO format and notice the dataloader output. I think it is the easiest way:)
For VG-150, I can provide the draft processing code (rebuilding the dataset based on the dataloader output of Neural Motifs). I haven't cleaned it (too busy right now), just for reference: https://drive.google.com/file/d/1vuT0OXla0fhx2ekcLoGs5YZp-VsrRLOM/view?usp=sharing
Hi, Dear authors of RelTR:
Your work is so solid that we are so impressed. Is it possible to release the code for preprocessing? so we could reproduce the results on other datasets.